[llvm-dev] Potentially unsafe loop optimization

Michael Kruse via llvm-dev llvm-dev at lists.llvm.org
Wed Feb 17 17:55:34 PST 2021


Am Mi., 17. Feb. 2021 um 18:39 Uhr schrieb Richard Kenner <kenner at adacore.com>:
> I don't think so because if %6 is loaded with the value 255 (-1), the
> branch won't go to loop.iter where the add is done.  So we never get
> to the add that generates poison.

Correct, I missed -1 ebing equal to 255 in i8.

Michael


More information about the llvm-dev mailing list